I.I Data cleaning code script
I cleaned the dataset, involving key metrics like flight and hotel discounts, using the 1.5x Interquartile Range (IQR) method to identify and exclude outliers. This approach ensured the dataset's diversity and integrity by removing extreme data points, retaining 43,566 transactions, and providing a solid foundation for further analysis.
I.II Data aggregation code script
In the analysis, I processed data from 5,988 customers, focusing on demographics, booking habits, and preferences. I calculated averages and ratios for key metrics like discounts and bookings, normalized variables for analysis, and aggregated session data to understand customer behavior comprehensively, ensuring insights into their travel preferences and demographic background.
II.Distribution consideration code script
In the perk distribution strategy, I focused on balancing cost control with user appeal, selectively offering high-cost perks such as exclusive discounts and waiving cancellation fees. I proposed low-cost, high-value rewards like free hotel meals for specific achievements and introduced a $20 credit for price-sensitive users composing the second-tier perk accomplishment reward system, ensuring efficient resource allocation and enhanced customer engagement.
III.I Perk distribution code script
In the perk distribution, I distributed perks according to user behavior and data insights.
III.II Perk achievement reward distribution
I assigned reward perks using K-means clustering, splitting users into two distinct groups based on their purchasing habits:
